Finance & Crypto

GitHub Copilot Shifts to Token-Based Pricing in Mid-2026

2026-05-01 20:45:29

Overview of the Billing Transformation

GitHub has announced that all Copilot plans will transition to a usage-based billing model starting June 1, 2026. This marks a fundamental shift from the current system of counting premium requests to a structure based on token consumption. The change aims to better align costs with actual usage and ensure long-term sustainability for the platform.

GitHub Copilot Shifts to Token-Based Pricing in Mid-2026
Source: github.blog

Why the Shift?

Copilot has evolved rapidly from a simple in-editor assistant into a sophisticated agentic platform. It now handles long, multi-step coding sessions, leverages the latest models, and iterates across entire repositories. Agentic usage—where the AI performs extended autonomous tasks—has become the default behavior for many users.

Under the old premium request model, a quick chat query and a multi-hour autonomous coding session were billed identically. GitHub had been absorbing the escalating inference costs of heavy usage, but this model proved unsustainable. The new usage-based pricing addresses this by charging proportionally to the computational resources consumed, helping maintain service reliability without imposing arbitrary caps on power users.

What’s Changing

Starting June 1, 2026, Premium Request Units (PRUs) will be replaced by GitHub AI Credits. These credits are consumed based on token usage—including input tokens, output tokens, and cached tokens—according to the published API rates for each model.

Key Details

Preparing for the Transition

To help users and administrators anticipate their costs, GitHub will release a preview billing experience in early May 2026. This tool will appear on the Billing Overview page of each user’s github.com account, offering visibility into projected credit usage before the June 1 switchover.

GitHub Copilot Shifts to Token-Based Pricing in Mid-2026
Source: github.blog

Additionally, GitHub recently implemented temporary adjustments to Copilot Individual plans (Free, Pro, Pro+, Student) and paused self-serve purchases of Copilot Business plans. These measures were taken to improve reliability and performance as the company prepares for the broader billing overhaul. Once the new system is in place, usage limits may be relaxed.

What This Means for Users

The move to usage-based billing reflects the growing complexity and power of AI coding tools. For users who rely on Copilot for occasional queries, costs are likely to remain stable—or even decrease—since they will only pay for what they use. For heavy users running extensive agentic sessions, the new model will more accurately reflect the computational resources consumed, potentially leading to higher bills but also providing a more sustainable service.

Administrators will gain better control through budget limits and clearer visibility into spending. The elimination of fallback models means that once credits are exhausted, usage will either stop or require additional credit purchases, depending on plan settings. This shifts the responsibility for managing consumption from GitHub’s arbitrary tiers to the user’s own needs and budgets.

Conclusion

GitHub’s transition to token-based billing for Copilot is a strategic response to the platform’s evolution and the rising costs of inference. By aligning pricing with actual usage, the company aims to ensure long-term reliability and fairness for all customers. Users should start reviewing their usage patterns now and plan for the new billing environment that takes effect on June 1, 2026.

Explore

Vercel Breach Exposes Danger of Third-Party OAuth Integrations: Experts Warn of 'Shadow AI' Sprawl Ireland Joins the Artemis Accords: A New Chapter in Space Exploration What You Need to Know About Most Frequently Asked Questions About Email Mark... Anthropic Launches Claude Opus 4.7 on Amazon Bedrock: Next-Gen AI for Enterprise Coding and Agents GitHub's Commitment to Reliability: Navigating Exponential Growth and Improving Availability